ChatGPT, Round Two

After sending some comments about my blog post on my initial excursions with ChatGPT earlier this week, I received a great deal of interesting follow-up. One colleague lamented that we may no longer be able to assess the knowledge of students, especially when teaching in online classes, because of their ability to easily generate answers to questions using systems like ChatGPT, especially if they could add to the text by Googling a few references to add to their answers.

I am not convinced of this, at least not with the use of ChatGPT, whose answers I noted were superficial, not completely correct, and not backed up with citations. Furthermore, asking for short answers to questions is hardly the only way to assess knowledge.

So to put ChatGPT to the next test, I pasted in, one by one, the 10 multiple-choice questions I give for the same unit of my introductory informatics course as the learning objectives of my first assessment.

 

While I have to give ChatGPT some credit for figuring out these were multiple-choice questions, it performed pretty poorly compared to most students in the course. It only answered 5 of the 10 questions correctly, and for two it got wrong, it did not even try to answer the questions, giving the excuse it could not browse the Internet (despite being trained on it!).

 

Some of the incorrect answers would be embarrassing even for a beginning informatics student to not know, such as which US government agency funds informatics training, that "meaningful use" was not part of the Affordable Care Act, and that a machine learning algorithm not involving clinicians does not adhere to Friedman's Fundamental Theorem.

 

The transcript of my interactions is pasted in below. The text entered into ChatGPT did not include the question numbers nor the correct answers indicated by asterisks. The paragraph following each question is how ChatGPT replied. My conclusion is that, for now, those of us who teach informatics in virtual settings do not (yet) have to worry about ChatGPT being used to cheat in our courses.

ChatGPT Transcript - December 10, 2022

1. Which part of biomedical and health informatics is most likely to address the critical informatics issues for the practice of a family physician?

a. Bioinformatics

*b. Clinical informatics

c. Clinical research informatics

d. Imaging informatics

e. Public health informatics

Clinical informatics is most likely to address the critical informatics issues for the practice of a family physician. This subfield of biomedical and health informatics focuses on the use of information technology to support the delivery of healthcare and improve patient outcomes. Clinical informatics encompasses a wide range of topics, including electronic health records, health information exchange, clinical decision support, and population health management. It is concerned with the design, development, and implementation of systems and tools to support the clinical work of healthcare professionals, such as family physicians.

2. Before the era of widespread computer use, the typical hospital physician spent what proportion of his or her time in activities away from the patient during office practice?

*a. 23-39%

b. 50%

c. 61-76%

d. 100%

I'm sorry, but I am not able to browse the internet to answer your question as I am a large language model trained by OpenAI. My knowledge is based solely on the text that I have been trained on, which has a cutoff date of 2021. I do not have any additional information beyond what I have been trained on.

3. Which entity of the US government funds research and training in biomedical and health informatics?

a. Agency for Healthcare Research & Quality

b. Center for Medicare and Medicaid Services

c. National Committee for Vital & Health Statistics

*d. National Library of Medicine

The National Institutes of Health (NIH) is the primary entity of the US government that funds research and training in biomedical and health informatics. The NIH is a federal agency that is part of the Department of Health and Human Services. It is the largest biomedical research agency in the world and provides funding for a wide range of research and training programs in the field of biomedical and health informatics. Some examples of NIH-funded programs in this area include the National Library of Medicine's Training and Education in Biomedical Informatics program and the Clinical and Translational Science Awards program, which support research and training in clinical informatics.

4. Before the emergence of the Affordable Care Act, a number of proposed healthcare reform plans for the US would provide for the establishment of a small number of highly regulated private insurance companies in a given region or state. This type of plan would be described by JR Reid as:

a. Beveridge

*b. Bismarck

c. National health insurance

d. Out-of-pocket

The type of healthcare reform plan described in the question, which involves the establishment of a small number of highly regulated private insurance companies in a given region or state, would be described by JR Reid as a Bismarck-type plan. Bismarck-type plans are named after Otto von Bismarck, the 19th century Prussian statesman who introduced the first national health insurance program in the world. Bismarck-type plans are characterized by a high degree of government control over the healthcare system, with the state playing a central role in regulating the provision of healthcare services and financing. In these plans, private insurance companies are typically required to operate within a regulatory framework set by the government, and may be subject to price controls and other forms of regulation. This type of plan is common in many European countries.

5. What proportion of US healthcare spending is devoted to prescription drugs?

*a. 10%

b. 20%

c. 33%

d. 50%

I'm sorry, but I am not able to browse the internet to answer your question as I am a large language model trained by OpenAI. My knowledge is based solely on the text that I have been trained on, which has a cutoff date of 2021. I do not have any additional information beyond what I have been trained on.

6. Compared to most other wealthy countries, the US has which of the following?

a. More hospital days

*b. More money spent

c. More nurses

d. More physicians

Compared to most other wealthy countries, the US has more money spent on healthcare. The US spends more on healthcare as a percentage of its gross domestic product (GDP) than any other country in the world. According to data from the World Health Organization, the US spent 17.2% of its GDP on healthcare in 2018, compared to an average of 9.6% for other high-income countries. This high level of spending is due in part to the relatively high cost of healthcare services in the US, as well as the large number of uninsured and underinsured individuals in the country.

7. Which of the following is not part of the original Affordable Care Act (ACA)?

*a. Incentives for meaningful use of electronic health records

b. Mandate that individuals buy health insurance

c. Outlawing of insurance denial due to pre-existing medical conditions

d. Programs to establish Accountable Care Organizations (ACOs)

Programs to establish Accountable Care Organizations (ACOs) are not part of the original Affordable Care Act (ACA). ACOs are groups of healthcare providers who come together to coordinate care for a specific group of patients, with the goal of improving the quality of care and reducing costs. ACOs were introduced as part of the Medicare program under the Patient Protection and Affordable Care Act of 2010, but were not part of the original ACA legislation. The ACA included a number of other provisions, such as incentives for the meaningful use of electronic health records, a mandate that individuals buy health insurance, and the outlawing of insurance denial due to pre-existing medical conditions.

8. About 50% of US healthcare spending is accounted for by which proportion of the population?

a. 1%

*b. 5%

c. 25%

d. 50%

Approximately 50% of US healthcare spending is accounted for by 5% of the population. This phenomenon, known as the "5/50 rule," refers to the fact that a small proportion of the population accounts for a disproportionate amount of healthcare spending. The 5/50 rule is often used to illustrate the unequal distribution of healthcare costs in the US, with a small group of individuals with chronic conditions or complex healthcare needs accounting for a significant portion of total spending. This pattern is driven in part by the high cost of treating certain conditions, such as cancer and heart disease, as well as the high cost of specialty drugs and other medical technologies.

9. Which of the following applications does not adhere to Friedman’s fundamental theorem of informatics?

*a. A machine learning algorithm that processes radiology images and determines the diagnoses more accurately than a physician

b. A telemedicine system that allows remote interaction via videoconference

c. A reminder system to administer cancer screening exams on an appropriate schedule

d. An information retrieval (search) system that helps the user select terms to improve search results

A telemedicine system that allows remote interaction via videoconference does not adhere to Friedman's fundamental theorem of informatics. The fundamental theorem of informatics, proposed by Edward H. Friedman, states that "the value of medical information is inversely proportional to the time and place of its creation and directly proportional to the time and place of its use." This theorem suggests that the value of medical information decreases as it becomes older and less relevant, and increases when it is used at the time and place where it is needed. A telemedicine system that allows remote interaction via videoconference does not necessarily adhere to this theorem, as it allows healthcare providers to access and use medical information from different times and places. An example of an application that adheres to the theorem would be a reminder system that administers cancer screening exams on an appropriate schedule, as it uses medical information at the time and place where it is needed to improve patient care.

10. A patient undergoing an appendectomy at a hospital near his suburban home is most likely receiving what type of healthcare?

a. Primary care

*b. Secondary care

c. Tertiary care

d. Quaternary care

A patient undergoing an appendectomy at a hospital near his suburban home is most likely receiving secondary care. Secondary care refers to healthcare services that are provided by specialized medical professionals, such as surgeons, specialists, and other healthcare providers. These services are typically provided in hospitals or other specialized medical facilities, and often require referral from a primary care provider. An appendectomy, which is a surgical procedure to remove the appendix, is an example of a medical procedure that is typically performed in a hospital or other specialized medical facility, and would be classified as secondary care. Primary care, on the other hand, refers to the first level of contact between patients and the healthcare system, and typically involves services such as preventive care, routine check-ups, and treatment of minor illnesses. Tertiary care and quaternary care are higher levels of specialized care that are provided for patients with complex or chronic conditions.

Initial Excursions with ChatGPT

Like many people, I found myself caught up this past week testing out ChatGPT, the new automated chat agent released by OpenAI. It seems like many people across the Internet are using it to test out its functioning in the kind of work they do, whether they are writers, programmers, or my fellow informatics researchers.

Naturally, I tested ChatGPT doing what I enjoy most, namely teaching. My first test was to enter the learning objectives for my well-known introductory biomedical and health informatics course, aka, 10x10 ("ten by ten"). Being the academic that I am, I was also curious what ChatGPT might report for my contributions to the informatics field. The transcript of my interaction is provided below, but first some commentary.

ChatGPT did a much better job with my learning objectives than its report of my accomplishments in informatics. In fact, it got quite a bit wrong about my own work. For example, it reported me as Editor of Journal of the American Medical Informatics Association (JAMIA) and Journal of Medical Internet Research ( JMIR), something my colleagues Sue Bakken and Gunther Eysenbach might dispute. It also gave erroneous information about books that I "authored." While I have edited and authored my textbook Health Informatics: Practical Guide, authored my textbook on information retrieval, and contributed chapters to many other books, I have no involvement whatsoever in the books listed by ChatGPT, Introduction to Health Informatics or Essentials of Health Information Systems and Technology. ChatGPT also listed first among my accomplishments in the field, "work on the development and implementation of EHRs." While I teach a great deal about the electronic health records (EHRs), it would be a stretch to say I am involved in their development and implementation. Clearly ChapGPT is an inferior alternative to Googling oneself.

For the learning objectives, ChatGPT did a little better, and gave reasonable, if incomplete, discussions on the topics in each. It still got some facts wrong, and also omitted what I might consider important details one might want to see in a succinct summary. Perhaps more importantly, ChatGPT provided no citations for what it said, which could allow a learner to check the interpretation or pursue more detail about the topic. I suppose the good news is that one still would do better to get more exhaustive text, figures, and citations covered in the 10x10 course (samples of which can be viewed at my site, informatics.health).

These are obviously early days for tools such as ChatGPT, and no doubt they will improve over time. But just as machine learning algorithms are not quite ready to replace human clinicians, they also are not quite ready to overtake other human tasks, such as teaching, at least not yet.

ChatGPT Transcript - December 10, 2022 (my text entry in italics)

Define biomedical and health informatics, the terms related to it, and its role in health, healthcare, public health, and biomedical research

Biomedical and health informatics is a field that focuses on the use of technology and information to improve the delivery of healthcare, public health, and biomedical research. It is an interdisciplinary field that combines computer science, data science, and healthcare to develop and use information systems and technology to manage, analyze, and interpret health-related data.

Some of the terms related to this field include electronic health records (EHRs), health information technology (HIT), health information management (HIM), and health informatics. These terms all refer to different aspects of using information and technology in the healthcare and public health sectors.

The role of biomedical and health informatics in healthcare is to improve the efficiency and effectiveness of care delivery by using technology and information systems to manage and analyze health-related data. This can include the use of EHRs to manage patient records, the use of HIT to monitor and manage the health of patients remotely, and the use of HIM to ensure the quality and security of health information.

In public health, biomedical and health informatics can be used to monitor and analyze population-level health data to identify trends and patterns, and to develop interventions and policies to improve the health of the population.

In biomedical research, this field can be used to manage and analyze large amounts of data from clinical trials, genetic studies, and other research to advance our understanding of health and disease.

Overall, the goal of biomedical and health informatics is to use technology and information to improve the quality, efficiency, and accessibility of healthcare, public health, and biomedical research.

Describe the major historical events of biomedical and health informatics

The field of biomedical and health informatics has a long and rich history, with many important events that have shaped its development. Some of the major historical events in this field include:

1. The development of electronic medical records (EMRs) in the 1950s, which marked the beginning of the use of computers in healthcare to store and manage patient information.
2. The establishment of the National Library of Medicine (NLM) in the United States in 1956, which became a major repository for health-related information and a driving force in the development of health informatics.
3. The creation of the first Health Information Management (HIM) program at the University of Minnesota in 1967, which marked the beginning of formal education and training in health informatics.
4. The development of the first clinical decision support system (CDSS) in the 1970s, which used computers to assist doctors in making diagnostic and treatment decisions.
5. The introduction of the first personal health record (PHR) in the 1980s, which allowed patients to manage their own health information and access it from any location.
6. The widespread adoption of the internet in the 1990s, which facilitated the sharing and dissemination of health-related information and the development of online health communities.
7. The passage of the Health Insurance Portability and Accountability Act (HIPAA) in 1996, which established national standards for the privacy and security of health information.
8. The growth of mobile health (mHealth) technologies in the 2000s, which enabled the use of smartphones and other devices to monitor and manage health-related information.

Overall, these and other historical events have played a significant role in the development and evolution of biomedical and health informatics as a field, and have had a profound impact on the way healthcare is delivered and research is conducted.

Discuss the major problems in healthcare motivating use of biomedical and health informatics

There are many challenges and problems facing the healthcare industry today that are motivating the use of biomedical and health informatics. Some of the major problems in healthcare that are driving the use of this field include:

1. The increasing cost of healthcare: Healthcare costs are rising rapidly, and many people are struggling to afford the care they need. Biomedical and health informatics can help to reduce these costs by using technology and data analysis to improve the efficiency and effectiveness of care delivery.
2. The growing burden of chronic diseases: Chronic diseases, such as diabetes, heart disease, and cancer, are becoming more prevalent and are a major cause of morbidity and mortality. Biomedical and health informatics can be used to monitor and manage these conditions, and to identify and intervene early to prevent their progression.
3. The shortage of healthcare workers: There is a shortage of healthcare workers in many parts of the world, which is putting a strain on the healthcare system and making it difficult for people to access care. Biomedical and health informatics can help to alleviate this shortage by enabling healthcare workers to work more efficiently and by supporting remote care delivery.
4. The need for better population health management: Many public health challenges, such as obesity, substance abuse, and infectious diseases, require a population-level approach to prevention and management. Biomedical and health informatics can be used to monitor and analyze population-level health data, and to develop and implement interventions to improve the health of the population.
5. The need for better data management and analysis: The healthcare industry generates vast amounts of data, but much of it is unstructured and difficult to manage and analyze. Biomedical and health informatics can help to organize and analyze this data to extract valuable insights and inform decision-making.

Overall, the major problems in healthcare are driving the use of biomedical and health informatics to improve the quality, efficiency, and accessibility of healthcare and to address some of the most pressing challenges facing the healthcare industry.

Compare and contrast the roles of various individuals in the health information technology workforce

The health information technology (HIT) workforce includes a wide range of individuals with different roles and responsibilities. Some of the major roles in the HIT workforce include:

1. Health information managers (HIMs): HIMs are responsible for the management and analysis of health-related data, including the design and implementation of health information systems and the development and maintenance of electronic health records (EHRs).
2. Health information technologists (HITs): HITs are responsible for the technical aspects of health information systems, including the design, development, and maintenance of software and hardware systems used in healthcare.
3. Clinical informaticists: Clinical informaticists are healthcare professionals with expertise in both medicine and health informatics. They use their knowledge of both fields to improve the quality and safety of care delivery, and to support the use of health information technology in clinical settings.
4. Health information exchange (HIE) coordinators: HIE coordinators are responsible for facilitating the exchange of health-related information between different healthcare organizations and systems. This includes ensuring the interoperability of different systems, and the security and privacy of health information.
5. Chief information officers (CIOs): CIOs are responsible for the overall management of information technology in healthcare organizations, including the development and implementation of information technology strategies and the supervision of IT staff.

Overall, the roles in the HIT workforce are diverse and varied, and individuals in these roles work together to support the use of health information technology in healthcare and public health.

Find and access the major sources of information for biomedical and health informatics in the scientific literature and on the World Wide Web

There are many sources of information for biomedical and health informatics in the scientific literature and on the World Wide Web. Some of the major sources of information for this field include:

1. PubMed: PubMed is a free database of biomedical literature maintained by the National Institutes of Health (NIH) in the United States. It contains millions of abstracts and full-text articles from scientific journals, and is a valuable resource for researchers and practitioners in biomedical and health informatics.
2. IEEE Xplore Digital Library: The IEEE Xplore Digital Library is a comprehensive collection of technical literature in electrical engineering, computer science, and related fields. It contains many articles and conference proceedings on topics related to biomedical and health informatics, and is a valuable source of information for researchers in this field.
3. Health Information and Libraries Journal: The Health Information and Libraries Journal (HILJ) is a peer-reviewed journal published by the Health Libraries Group of the Chartered Institute of Library and Information Professionals (CILIP) in the United Kingdom. It covers a wide range of topics related to health information and health informatics, and is a valuable source of information for librarians and other professionals working in this field.
4. Journal of the American Medical Informatics Association (JAMIA): JAMIA is a peer-reviewed journal published by the American Medical Informatics Association (AMIA). It covers a wide range of topics related to health informatics, including the use of technology and information to improve the delivery of healthcare and public health.
5. Health IT Buzz: Health IT Buzz is a blog maintained by the Office of the National Coordinator for Health Information Technology (ONC) in the United States. It provides news and updates on health information technology and health informatics, and is a valuable source of information for those interested in this field.

Overall, these and other sources of information are valuable for researchers and practitioners in biomedical and health informatics, and provide a wealth of information on the latest developments and trends in this field.

What are the contributions of William Hersh to informatics?

William Hersh is a well-known figure in the field of informatics. He is a professor and the chair of the Department of Medical Informatics & Clinical Epidemiology at the Oregon Health & Science University School of Medicine.

Some of Hersh's notable contributions to the field of informatics include his work on the development and implementation of electronic health records (EHRs), the use of natural language processing for clinical text analysis, and the integration of clinical research data with EHRs. He has also contributed to the development of informatics curricula and has written extensively on the subject.

Hersh is the author of several books on informatics, including "Introduction to Health Informatics" and "Essentials of Health Information Systems and Technology". He is also the editor-in-chief of the Journal of the American Medical Informatics Association and the Journal of Medical Internet Research.

Overall, Hersh has made many important contributions to the field of informatics, including his work on the development and implementation of EHRs, the use of natural language processing for clinical text analysis, and the integration of clinical research data with EHRs.

The Most Profound Event in My Informatics Career? Unquestionably the HITECH Act

I recently took part in a small workshop exploring the benefits and challenges for artificial intelligence in medicine. Many of the participants were innovative medical educators, and most of them were still practicing clinical medicine. Although the workshop covered a broad range of topics, one theme stood out to me, perhaps because of my role as an academic informatician. In a nutshell, when we arrived at the topic of implementing AI systems in clinical practice, many laments were expressed about their impact on clinicians and insuring that the use of AI "cannot be as awful as the EHR."

As one who has been working in biomedical and health informatics for over three decades, this discussion about impact on practice and the experience of the electronic health record (EHR) drove home that unquestionably the most profound event in informatics during my entire career has been the the Health Information Technology for Economic and Clinical Health (HITECH) Act. Our field was changed substantially with the passage of the HITECH Act as part of the American Recovery and Reinvestment Act (ARRA) in 2009, leading our present world to be substantially different from the world before it. I also find that many who lament about the negative impacts of the EHR on healthcare are not aware of the history of HITECH.

Those of us who have been around long enough can remember informatics before HITECH. While EHRs had been adopted by many healthcare organizations, including my own, overall use was modest. We were also encouraged, especially in the decade of the 2000s before HITECH, by the growing amount of research published showing that the EHR and associated clinical decision support functionality led to error reduction, cost savings, and improved outcomes. Indeed, a systematic review in 2006 summarized the mostly positive studies, although noting they mostly came from "health IT exemplar" institutions [1]. An updated review in 2009 found similar results [2]. My own view at the time was captured from my participation in a panel on the National Public Radio (NPR) show, Talk of the Nation - Science Friday, in 2005.

Another commentary published in 2010 noted that four "tribes" of healthcare improvement - focused on quality improvement, payment reform , consumer engagement, and health IT - were coalescing in the run-up to health care reform under President Barack Obama, for which HITECH was described by some as a "down payment" [3]. And the pre-HITECH enthusiasm all reached a boil when the ARRA legislation included the HITECH Act and its $30+ billion for incentives for EHR adoption [4].

Here is how I concluded a post in this blog from early 2010 highlighting the details of the HITECH Act:

This is a defining moment for the informatics field. Never before has such money and attention been lavished on it. HITECH provides a clear challenge for the field to 'get it right.' It will be interesting to look back on this time in the years ahead and see what worked and did not work. Whatever does happen, it is clear that informatics lives in a HITECH world now.

To many in informatics, the HITECH Act may now be a distant event, and some in the field who are younger may not even had yet started working in it. But clearly the informatics world is much different now. Informatics is certainly more prominent in operational healthcare settings. Ever increasing numbers of healthcare organizations have anointed clinical leadership devoted to informatics [5].

In retrospect, the widespread EHR adoption of HITECH was carried out too hurriedly, and glossed over some major problems, not the least of which was inadequate adherence to data standards and interoperability. Indeed, it can be said that the health IT components of the 21st Century Cures Act, passed with a bipartisan majority in 2016, were some necessary corrections to the problems emanating from the HITECH Act [6]. The Cures Rule enshrining SMART on FHIR for interoperability and information blocking prohibitions to facilitate movement of data can be viewed from that perspective.

Whether deserved or not, the informatics community is included in the blame for the facts that EHRs slow clinicians down [7, 8], fill the record with questionable and duplicative information [9], and contribute to clinician burnout [10, 11]. The informatics field should accept part of the blame but must also be part of the solution. Actions such as the AMIA-led 25x5 initiative to reduce the burden of EHR documentation to 25% of its current level within five years are important. I also believe there are technologies that will improve the capture of data and clinical thought processes into the EHR, such as voice recognition and reworking of the user interface to facilitate care rather than just documenting it.

Clearly the HITECH Act was the most profound event of our field in modern times. The informatics world before and after HITECH is very different, for better or worse. Ever the informatics optimist, there is great opportunity to improve the use of EHR and other systems to deliver modern healthcare. I am confident we will rise to the challenge.

References

1. Chaudhry, B., Wang, J., Wu, S., Maglione, M., Mojica, W., Roth, E., Morton, S.C., Shekelle, P.G., 2006. Systematic review: impact of health information technology on quality, efficiency, and costs of medical care. Ann Intern Med 144, 742–752.

2. Goldzweig, C.L., Towfigh, A., Maglione, M., Shekelle, P.G., 2009. Costs and benefits of health information technology: new trends from the literature. Health Aff (Millwood) 28, w282-293.

3. McKethan, A., Brammer, C., 2010. Uniting the tribes of health system improvement. Am J Manag Care 16, SP13-18.

4. Blumenthal, D., 2011. Implementation of the federal health information technology initiative. N Engl J Med 365, 2426–2431.

5. Kannry, J., Sengstack, P., Thyvalikakath, T.P., Poikonen, J., Middleton, B., Payne, T., Lehmann, C.U., 2016. The Chief Clinical Informatics Officer (CCIO): AMIA Task Force Report on CCIO Knowledge, Education, and Skillset Requirements. Appl Clin Inform 7, 143–176.

6. Kesselheim, A.S., Avorn, J., 2017. New “21st Century Cures” Legislation: Speed and Ease vs Science. JAMA 317, 581–582.

7. Tutty, M.A., Carlasare, L.E., Lloyd, S., Sinsky, C.A., 2019. The complex case of EHRs: examining the factors impacting the EHR user experience. J Am Med Inform Assoc 26, 673–677.

8. Zheng, K., Ratwani, R.M., Adler-Milstein, J., 2020. Studying Workflow and Workarounds in Electronic Health Record-Supported Work to Improve Health System Performance. Ann Intern Med 172, S116–S122.

9. Steinkamp, J., Kantrowitz, J.J., Airan-Javia, S., 2022. Prevalence and Sources of Duplicate Information in the Electronic Medical Record. JAMA Netw Open 5, e2233348.

10. Gardner, R.L., Cooper, E., Haskell, J., Harris, D.A., Poplau, S., Kroth, P.J., Linzer, M., 2019. Physician stress and burnout: the impact of health information technology. J Am Med Inform Assoc 26, 106–114.

11. Murthy, V., 2022. Health Worker Burnout — Current Priorities of the U.S. Surgeon General [WWW Document]. https://www.hhs.gov/surgeongeneral/priorities/health-worker-burnout. 

A New Pathway For My Career in Biomedical and Health Informatics

This year, 2022, marks a quarter-century of my leadership of academic informatics at Oregon Health & Science University (OHSU). I have served for 19 years as the one and only Chair of the Department of Medical Informatics & Clinical Epidemiology (DMICE) in the OHSU School of Medicine. I served six years before that as Head of its predecessor free-standing Division of Medical Informatics & Outcomes Research, which was established in 1997.

I now plan to step down as Chair of DMICE while continuing at OHSU as a Professor in DMICE, focusing my activities on research, education, writing, and mentoring. I will support the new Chair and prioritize the continued success of the department.

I am pleased to hand off the department at a time when it is doing well academically and financially. The faculty are productive and well-funded in their research, and our educational programs continue to attract strong enrollment and prepare students for diverse careers in our field.

My own research and educational activities are also productive and well-funded. I plan to continue my research in applying information retrieval (IR, also known as search) methods to tasks such as patient cohort discovery though data from the electronic health record (EHR) for which I am funded with an R01 grant from the National Library of Medicine. I also intend to continue working in the evaluation of machine learning for clinical applications, mainly focused on diagnosis of rare diseases through EHR data.

I will also continue teaching both in the OHSU Biomedical Informatics Graduate Program as well as in external collaborations, such as with the American Medical Informatics Association (AMIA) in the 10x10 ("ten by ten") program and the Clinical Informatics Board Review Course (CIBRC). I will also continue existing and new National Institutes of Health (NIH)-funded educational activities, including our recently renewed NLM T15 Training Grant, a new NLM R25 establishing a new college summer internship for students historically underrepresented in our field, the Data Science Initiative for Africa in collaboration with the University of Cape Town, and the new Bridge2AI Program.

Arrival of 8th Edition of Health Informatics: Practical Guide

I am pleased to announce publication of the 8th Edition of Health Informatics: Practical Guide. The goal of this textbook is to provide a comprehensive introduction to the field of health informatics (also known as biomedical and health informatics or just plain informatics) for students. The book can also serve as a current update for those who are already practitioners. As in any professional field, there is a knowledge base and history of scientific progress in informatics that informs current practice, and this book aims to impart that.

The impetus for this book came from Robert Hoyt, MD, who served as its lead Editor for the first seven editions. In preparing the seventh edition, Dr. Hoyt engaged me as a Co-Editor, with the plan for me to become the Editor of the 8th edition. Dr. Hoyt and his wife Ann Yoshihashi, MD have been great mentors in the process of putting a book like this together. As with the previous editions, the book is available in print and eBook formats.

This book is aimed at those who wish to learn about the field of biomedical and health informatics. The content is almost completely redone from previous editions and roughly follows a course I have taught for three decades to about 5000 students in a graduate course at Oregon Health & Science University, the American Medical Informatics Association 10x10 ("ten by ten") program, and a number of other derivative courses for medical students, graduate students, and health IT professionals. I have engaged a variety of other authors for many of the chapters to give additional expertise and perspective. This edition of the book is completely rewritten from earlier editions, reflecting the profound change and progress in the field.

I have established a Web site for the book, which has links to purchasing the book in print and eBook formats from the publisher, Lulu.com. The book is also available from Amazon.com in print and Kindle formats. The Web site also contains sample chapters in PDF and EPUB formats, and will maintain an Errata page of errors discovered in the book.

The chapters of the book include:

1. Introduction to Biomedical and Health Informatics
2. A Short History of Biomedical and Health Informatics
3. Computing Concepts for Biomedicine and Health
4. Electronic Health Records
5. Standards and Interoperability
6. Data Science and Artificial Intelligence
7. Clinical Decision Support
8. Natural Language Processing
9. The Role of Health Informatics in Safety, Quality, and Value
10. Health Information Exchange (HIE)
11. EHR System Selection and Implementation
12. Telemedicine and Telehealth
13. Privacy and Security
14. Information Retrieval (Search)
15. Clinical Research Informatics
16. Translational Bioinformatics
17. Nursing Informatics
18. Consumer Health Informatics
19. Public Health Informatics
20. Evidence-Based Medicine
21. Imaging Informatics
22. Ethical Issues in Health Informatics
23. Human-Computer Interaction

Reflections on Publishing a Negative Study

Although scientists are supposed to be disinterested observers of the results of their work, the reality is that none of us really want to see a negative result from our research. This is especially the case in fields like biomedical and health informatics, where we aim to develop tools that can be used to help patients and/or clinicians with their health or for delivery of care.  Nonetheless, it is critically important to publish such outcomes, especially if the research was methodologically sound. This is especially the case in currently hyped technologies such as machine learning.

One of the challenges for negative studies is the myriad of potential reasons why they failed, which sometimes makes the true reason hard to pinpoint. Was it poor methods, inadequate data, or mis-application of the technology? Or more than one of those? But especially in the case of studies that involve human subjects who may be exposing themselves to risk, we must always remember the words of noted physician-scientist Iain Chalmers, who stated that failure to publish studies with patients is a form of scientific misconduct.(1)

A couple years ago, my colleague Aaron Cohen and I received some funding from a pharmaceutical company that had developed a highly effective new treatment for a rare disease, acute intermittent porphyria (AIP).(2) The new treatment, although expensive, significantly reduces the severe and sometimes disabling manifestations of AIP. As such, tools to identify patients who might have the disease could significantly improve their lives. Our approach used data from their electronic health record (EHR).

We were thrilled when our machine learning model was found to identify a group of patients with the classic presentation of AIP, yet the diagnosis had never been considered.(3) AIP is known to be one of those rare diseases that often goes undiagnosed for long periods of time. Diagnosis of these types of rare diseases is considered to be a major use case for machine learning in healthcare. From a collection of over 200,000 patient records in the research data warehouse at our institution, we manually reviewed the top 100 ranking patients identified by the model. Twenty-two of these patients were determined to have the manifestations of AIP but with no occurrence of the string "porph" anywhere in their record, i.e., not in lab tests, notes, or diagnoses. As the test to diagnose AHP is a relatively simple and inexpensive urine porphobilinogen test, we developed a clinical protocol to invite such patients for testing.

We encountered a number of challenges in developing and implementing the clinical protocol. For example, what is the role of the patient's primary care or other provider in the decision to offer testing? Our institutional review board (IRB) determined that the process should a decision by the provider, who would then allow us to contact the patient to offer testing. Once given approval by the provider, we then had to convince the patients, many of them skeptical of the healthcare system long unable to diagnose their symptoms, to come in for testing. Another challenge in implementing the protocol was that it took place during the COVID-19 pandemic, when many people wanted to avoid healthcare settings, though by the time we were recruiting patients, the COVID-19 vaccines had started to become available.

As a result, we could only convince 7 of the 22 patients to undergo the simple and free test we were offering them. And as noted in the paper, none of them were positive.(4) Because these patients did submit to a clinical protocol, and because we believe that any attempt to clinically validate machine learning is important, we decided to submit our results to a journal as a Brief Communication, i.e., not the definitive study but clearly worth reporting. The report has now been published.

Clearly there are many possible reasons for our failure to diagnose any new cases of AIP. Not the least of these is the fact that AIP is a rare disease, and there may have been few if any new cases to be found. Certainly if we had more resources or time, we could have invited for testing more people beyond the first 100 cases we evaluated who may also have been found to have the classic presentation yet for whom the diagnosis was never considered. It is also possible that our machine learning model could have been better at identifying true cases of the disease. There may have been confounders in that those actually coming for testing were not the most likely to have the diagnosis.

Could this study have benefited from proactive genotyping? There may be a role for gene sequencing in this situation, but AIP is a condition of incomplete penetrance, i.e., just because one has the genotype does not mean they will develop the manifestations of the disease. In other words, genotyping would not be enough.

My main hope is that this contribution shows that we should be pursuing such studies aiming to apply machine learning in real-world clinical settings. As one who has been critical that machine learning has not been "translational" enough, I hope there will be a role for larger and more comprehensive studies of many different uses that have been found to be beneficial in model-building studies.

References

1. Chalmers, I., 1990. Underreporting research is scientific misconduct. JAMA 263, 1405–1408.

2. Balwani, M., Sardh, E., Ventura, P., Peiró, P.A., Rees, D.C., Stölzel, U., Bissell, D.M., Bonkovsky, H.L., Windyga, J., Anderson, K.E., Parker, C., Silver, S.M., Keel, S.B., Wang, J.-D., Stein, P.E., Harper, P., Vassiliou, D., Wang, B., Phillips, J., Ivanova, A., Langendonk, J.G., Kauppinen, R., Minder, E., Horie, Y., Penz, C., Chen, J., Liu, S., Ko, J.J., Sweetser, M.T., Garg, P., Vaishnaw, A., Kim, J.B., Simon, A.R., Gouya, L., ENVISION Investigators, 2020. Phase 3 Trial of RNAi Therapeutic Givosiran for Acute Intermittent Porphyria. N Engl J Med 382, 2289–2301.

3. Cohen, A.M., Chamberlin, S., Deloughery, T., Nguyen, M., Bedrick, S., Meninger, S., Ko, J.J., Amin, J.J., Wei, A.J., Hersh, W., 2020. Detecting rare diseases in electronic health records using machine learning and knowledge engineering: Case study of acute hepatic porphyria. PLoS ONE 15, e0235574.

4. Hersh, W.R., Cohen, A.M., Nguyen, M.M., Bensching, K.L., Deloughery, T.G., 2022. Clinical study applying machine learning to detect a rare disease: results and lessons learned. JAMIA Open 5, ooac053.

2022 Update of Informatics.Health

I am pleased to announce an update of my Web site that provides an introductory overview of biomedical and health informatics. Entitled, What is Biomedical & Health Informatics?, I created this site two decades ago to provide an answer to that question I used to be asked (less so in modern times). I still maintain and keep the site up to date both to continue to provide an overview of the field as well as demonstrate a portfolio of some of the learning technology we use in our virtual courses in the Biomedical Informatics Graduate Program at Oregon Health & Science University (OHSU).

In recent years I was able to secure the domain name, Informatics.Health.

With the 2022 updating of my larger course that is offered in the American Medical Informatics Association (AMIA) 10x10 ("ten by ten") program, I have now updated the content of the WhatIs site. The main part of the site is the 11 lecture segments on the following topics:

• What is Biomedical and Health Informatics? (1) (24:32)
• What is Biomedical and Health Informatics? (2) (18:49)
• A Short History of Biomedical and Health Informatics (22:30)
• Resources for Field: Organizations, Information, Education (25:29)
• Clinical Data (15:08)
• Examples of the Electronic Health Record (EHR) (24:56)
• Data Science and Artificial Intelligence (1) (14:15)
• Data Science and Artificial Intelligence (2) (22:07)
• Information Retrieval (Search) (23:18)
• Information Retrieval Content (29:09)

The lectures can be viewed on just about any Web platform, and work fine on mobile devices. The site also contains links to books, articles, organizations, and educational Web sites. The materials on the site are freely available and have been used by many educators and others.

3000 by 2022 - A New Milestone for the 10x10 Course

One of the most enjoyable and impactful activities of my career has been the 10x10 ("ten by ten") course, which is an online introductory course in biomedical and health informatics that I teach in partnership with the American Medical Informatics Association (AMIA). The course recently reached a new milestone, surpassing 3000 people completing the course since its inception in 2005. My course from Oregon Health & Science University (OHSU) was the original and is the largest course of the AMIA 10x10 Program.

The OHSU 10x10 course provides an up-to-date detailed introductory overview of biomedical and health informatics. The content delivery is online, followed by an optional in-person session, typically at an AMIA meeting, at its conclusion. About 10-15% of those completing the course take the optional final exam and pursue further study in the field, usually but not exclusively at OHSU.

The 10x10 program gets its name from the program's original goal in 2005 to train 10,000 people in biomedical and health informatics by the year 2010. The goal for 10,000 people came from a stated need by then-President of AMIA Dr. Charles Safran that there should be at least one physician and one nurse trained in informatics in each of the approximately 6000 hospitals in the US. I adapted an already-existing online introductory informatics course and we set the goal of 10,000 people completing the course by 2010, leading me to devise the moniker, 10x10. Even though 10,000 people did not come forward to take the course by 2010, about 1000 people did complete the OHSU offering by 2010.

 

After 2010, the 10x10 program continued to attract enrollment. The positive feedback of those completing the course has always been gratifying. Now, a new milestone has been reached with the cohort who recently completed the course, surpassing a total of 3000 individuals having completed the OHSU 10x10 course since 2005. A total of 111 offerings of the OHSU course have been provided since that time. While most of the course enrollment has come from the US, several international organizations have partnered to offer the course based in their countries, including Gateway Consulting (Singapore), King Saud University (Saudi Arabia), the Ministry of Health (Israel), and Abu Dhabi Health Services (SEHA, United Arab Emirates). There have also been partnerships to offer the course with domestic organizations in the US, including the American College of Emergency Physicians, the Academy of Nutrition and Dietetics (AND), the New York State Academy of Family Practice, the Mayo Clinic, the Centers for Disease Control and Prevention, and others. The table below shows the course partners, the number of offerings with them, and the number of people completing the course.

 

 

The figure below shows the enrollment by year. The largest amount of enrollment came during the heyday of the "meaningful use" era in the early 2010s, dipped somewhat later in the decade, but has continued strong since then.

 

Due to the online delivery of the course, it had no trouble continuing during the COVID-19 pandemic, although the in-person sessions at the end of the course were transitioned to virtual delivery. With the receding of the pandemic, those sessions have returned now to an in-person format. Of course, the content of the course was affected by the pandemic and its impacts on the field. However, the more fundamental topics of the field, including electronic health records, standards and interoperability, machine learning and artificial intelligence, and others have still been covered in a comprehensive and up-to-date manner.

The course will continue to be offered beyond this milestone, and in fact a new offering is beginning this month. In-person sessions for those completing the course will take place at upcoming AMIA meetings this year, including the Clinical Informatics Conference (Houston, TX - May 23-26) and the Annual Symposium (Washington, DC - November 5-9).

As the OHSU agreement with AMIA is non-mutually-exclusive (i.e., AMIA can offer other 10x10 courses and OHSU can use the course materials in other courses), the content of the OHSU 10x10 course has been delivered to students under other course names. The other major use of the content is in the course, BMI 510/610, which is the introductory course in the Health & Clinical Informatics Major of the OHSU Biomedical Informatics Graduate Program. This course has been offered since the OHSU graduate program was launched in 1996 and every academic quarter since 2000. After the most recent academic term, 1598 people have completed this instance of the course.

The course content was also converted into a virtual two-week block course for medical students at the onset of the pandemic. The course was made available to medical students within and outside OHSU, with 44 OHSU medical students participating in three offerings and 178 medical students from outside OHSU participating in eight offerings during 2020. This block course continues to be offered to OHSU medical students.

 

Other offerings of the course have been provided to health systems, including Kaiser Permanente Northwest, Providence Health and Systems, and Bangkok (Thailand) Hospital. The course has also been offered to H3ABioNet, the Pan African Bioinformatics Network for the Human Heredity and Health in Africa (H3Africa) consortium, of which I serve on the Scientific Advisory Board.

 

The course  also inspired federal legislation in the US, leading to the 10,000 Trained by 2010 Act, introduced by Rep. David Wu (D-OR). The bill was passed by the US House in the 111th Congress (2009-2010), and elements of it were incorporated into the HITECH Act. The HITECH program included funding for Health IT Workforce Development, including grants awarded to OHSU.

Probably the most gratifying aspect of the course is the feedback I get from students who appreciate the broad but sufficiently deep content, i.e., getting the big picture while also learning the rationale behind it. I enjoy keeping in touch with those who have taken the course, especially when I run into them in airports, at conferences, or other places. I also value the motivation that the course provides to me in keeping up with advances across the entire field of biomedical and health informatics.

Clinical Informatics Subspecialty Practice Pathway Extended for Three Additional Years

A three-year extension to the Practice Pathway of board certification eligibility for the clinical informatics (CI) subspecialty has been approved by the American Board of Medical Specialties (ABMS) for the American Board of Preventive Medicine (ABPM). This is the second extension of the so-called "grandfathering" pathway that now allows physicians with a primary boarded specialty to achieve board eligibility without formal fellowship training through 2025. This means that physicians who desire to become board-certified in CI will be able to qualify to sit for the board examination by time working in the field or completing "non-traditional" training, the latter which may include a master's degree from a "24 month Masters or PhD program in Biomedical Informatics, Health Sciences Informatics, Clinical Informatics, or a related subject from a university/college in the US and Canada, deemed acceptable by ABPM (e.g. NLM university-based Biomedical Informatics Training)," such as the online Master of Science Program at Oregon Health & Science University (OHSU).

The CI field has expressed mixed feelings on this extension. In particular, a group of CI fellowship Program Directors published a letter in the journal Applied Clinical Informatics (ACI) opposing the extension, noting that the time for grandfathering has passed, and extending the practice pathway will adversely impact pursuit of fellowships.[1] I rebutted this letter, arguing that the field must find alternatives to two-year in-place fellowships to allow broader entry into the field.[2] In particular, a two-year in-place fellowship may place undue burdens on those who wish to become board-certified in CI long after they completed their primary training and would be unable to uproot from job, family, and/or other obligations.

Here are some snippets from my ACI letter arguing for alternatives to in-place fellowships:

I agree that we have passed the point where the Practice Pathway should allow physicians to become board-certified with essentially no formal training. However, I argue instead for this approach to be transformed into a method by where those who are unable to halt careers, salary, and family to pursue a pathway to certification that is mostly virtual and asynchronous yet still rigorous and supervised. Ironically, the pandemic has taught us that CI practice and education can be carried out in a mostly virtual format.

The in-place model for fellowship training made sense in the 20th century model of career development, where one completed education and training in their chosen profession and then entered the workforce for their career. In the 21st century, however, many professionals, especially in knowledge careers, change career pathways long after their primary education and training experience.

I oppose CI fellowships being completely remote, but it would be novel and innovative if there were some sort of hybrid training pathway, with fellows connected to an institution that could offer courses and allow supervised, mentored training experiences in healthcare organizations. Fellows would participate in a mostly remote way, but also have periodic in-person experiences, including stints that might be for several weeks or more and would involve direct interaction with faculty and colleagues. The field of Hospice and Palliative Medicine developed such an approach prior to the COVID-19 pandemic. Even CI somewhat emulates this approach now, as a half-dozen CI fellowships make use of online didactic courses from OHSU.

I applaud that for now the Practice Pathway will still allow those to pursue board certification. Hopefully the CI field can transition to a training process beyond the Practice Pathway that allows entry into the field without an in-place fellowship. As informaticians, we should be at the forefront of pioneering this approach in graduate medical education.

What are eligibility requirements for the CI subspecialty? They are essentially unchanged from the last time I posted about them in this blog in 2019, with the exception that the "grandfathering" pathway is now available through 2025.

References

1. Turer, R.W., Levy, B.P., Hron, J.D., Pageler, N.M., Mize, D.E., Kim, E., Lehmann, C.U., 2022. An Open Letter Arguing for Closure of the Practice Pathway for Clinical Informatics Medical Subspecialty Certification. Appl Clin Inform 13, 301–303. https://doi.org/10.1055/s-0042-1744386.

 

2. Hersh, W.R., 2022. The Clinical Informatics Practice Pathway Should Be Maintained for Now but Transformed into an Alternative to In-Place Fellowships. Appl Clin Inform 13, 398–399. https://doi.org/10.1055/s-0042-1745722.

Receding of the Pandemic: Will the Third Time Be the Charm?

Today, the Governor of Oregon is lifting the state's indoor mask mandate and ending the state's public health COVID-19 emergency. Like most US states, Oregon had a large Omicron wave of cases, hospitalizations, and deaths, although as through all of the pandemic still far below US national averages. At Oregon Health & Science University (OHSU), the number of patients in the hospital and ICU continues to fall each day. The mask mandate in non-healthcare buildings at OHSU will be dropped tomorrow and my department will return to some activities in-person with the start of the spring quarter at the end of March.

Will the pandemic finally recede now and allow us to return to an albeit new normal? We have been down this road before. The first came in the late spring and early summer of 2021. Vaccination had become relatively widespread, and the large wave of hospitalizations and mortality from late 2020 and early 2021 appeared to be subsiding. In Oregon, all mask and other public health mandates were lifted, and life seemed to be returning to normal.

 

Sadly, however, the Delta wave started in the late summer of 2021 and dashed hopes that the fall would see a return to relative normal. As the Delta wave subsided in the late fall, a second era of opportunity seemed to be coming again. Although somewhat more muted than the first reprieve, it looked as if a modified normal might occur in early 2022.

But then, of course, Omicron came, and with it a new wave of hospitalizations and mortality. One fortunate aspect of the Omicron wave is that there has been clear evidence that vaccination provides protection. Even if not preventing SARS-CoV-2 infection completely, vaccination does appear to limit the worst of the infection for most people.

Now the Omicron wave is receding, and hopefully the worst of the pandemic with it. Although some might be gun-shy to feel optimistic, it is clear that there can be a path to living with the virus and a new approach that balances public health requirements with individualizing risk. I look to those physician experts who express cautious optimism and allowing of individual decision-making within the extremes of those at the ends of the spectrum. My favorites include Ashish Jha, Bob Wachter, Leana Wen, and of course the prolific Eric Topol. I also believe that the federal government's new COVID-19 plan is reasonable, with its emphasis on vaccination, testing, and treatment. This is especially the case with new oral anti-viral treatments shown to be highly effective.

The most unfortunate aspects of this pandemic has been its politicization, it becoming part of the culture wars in the US. I am not unsympathetic to those who want to move on. I do not particularly enjoy wearing masks, even though I do so and will continue doing so when it is necessary to protect myself or others. It saddens me that one of the most speedy and effective vaccines ever developed is being met by so much resistance. In addition, the manipulation and misinformation is saddening for a medium I always hoped would lead to dissemination of communication and knowledge across humanity. Wishful thinking, I suppose.

I find particularly sad the misunderstanding of science and the gotcha politics of when knowledge changes based on new research. One of the best quotes comes from Mohamad Safa, an environmental activist who stated on Twitter: "Science is not truth. Science is finding the truth. When science changes its opinion, it didn't lie to you. It learned more."

As the pandemic recedes, what will be my approach? Being relatively healthy and in my early 60s, I would probably weather a COVID-19 infection reasonably well. I will not go out looking for one, but I also will not have dire fear of getting one. One irony about the pandemic period is that I have not been infected so much as a cold, which I normally get once to twice per year. I will also respect the public health concerns for COVID-19. I will not hesitate to wear a mask when I am asked to do so, especially when it protects someone who might be at higher risk of complications from infection. I will also continue to join the chorus of those advocating for more vaccine equity across the world.

The months ahead will be a large natural experiment in the United States, as mandates are lifted. We will see whether the collective immunity we have achieved through vaccination and natural infection will be enough to keep SARS-CoV-2 under control, especially those at highest risk of complications from infection.

Considerations and Transitions for Narrating Lecture Slides

For over two decades, a significant portion of my teaching at Oregon Health & Science University (OHSU) has been online, and one of the main modalities I have used is narrated lectures, in particular, voice-over-Powerpoint slides. I know that many bemoan the use of Powerpoint for any presentations, including teaching. However, I find slides a very useful way to organize information, utilize simple graphics, and develop a big picture for the knowledge I aim to convey. As I use my teaching slides for other purposes, such as giving talks, another critical need for me is to maintain the source Powerpoint files.

This makes my choice of tool(s) to create voice-over-Powerpoint files critically important in my work. Over the years, I have used different tools for doing so. My initial foray used a technology that some Internet long-timers might remember, called RealMedia. Another imperative in the early days, circa 1999, was that the files not be too large so as not to require substantial bandwidth, since those were the early days of broadband Internet, and many people were still connecting, especially from home, via telephone modems.

Another lesson learned in those early days was that learners preferred online lectures to be broken down into 15-20 minute segments. This also made recording easier, as one did not have to record an entire lecture in a single session.

Over time, there was development of new tools. RealMedia eventually fell out of favor with many users and developers, and for a while I used a now-defunct tool from Adobe called Presenter. It used an output format that was very popular for many years, Flash, which was retired a few years ago.

In 2007, I discovered another tool that served me well for a long time, Articulate Presenter. One of its key features was the ability to record a single slide at a time. All of the tools I used before then required the lecture over the whole slide deck to be recorded in a single session. Being able to record slide by slide was useful for several reasons, not the least of which was that most of my updating of curricular content involved incremental update of single slides and not the entire lecture segment. In other words, a small part of a topic might change, and by being able to update individual slides, I could update just those slides than needed it.

Another feature I valued from Articulate Presenter was being able to configure the output with a navigation and notes frame on the left, enabling individuals slides to navigate to specific slides, and for the text in the notes field of the slides to be viewed. The output was generated in Flash when it was in its heyday. and then transitioned to HTML 5 as Flash was being retired.

But there is one big limitation of Articulate Presenter, which is that it only runs on Microsoft Windows. My primary computer when I started using Articulate Presenter was a Windows PC, but in the early 2010s, I was feeling the pull to return to the Macintosh platform that I had moved away from in the mid-1990s. Part of the allure was that Apple itself was transitioning to Intel chips, meaning that Macintosh computers could run Windows and its applications.

I made the transition back to the Mac in 2012, and fairly seamlessly moved my Articulate Presenter work to running Windows on Parallels. But although the solution worked, it was somewhat clunky, and having to launch Windows to do any work in Articulate Presenter required a number of steps. Despite the protestations of myself and many other users (evidenced in their support forums), Articulate never developed a native Mac version of Presenter.

In the last couple years, of course, Apple has transitioned away from Intel chips. This is a good thing for Mac users generally, as the speed and power consumption is remarkable. And while there is supposedly a version of Windows that runs on the ARM chips that Apple is now using, it does not run most native applications, including Articulate Presenter.

Fortunately during this same time, Microsoft has been improving the slide narration capabilities, and one can now easily narrate slide by slide directly in Powerpoint. And then the output can be exported to a standard video format such as MP4 or MOV. In addition, OHSU uses the Echo 360 platform, which allows video files to be posted to a server. While the video output of Powerpoint does not have all the navigation features of Articulate Presenter, the learner can still navigate around the lecture. Echo 360 also offers automated closed-captioning, although it is far from perfect.

This transition will require me to re-record all lectures completely, but many lectures are due for that anyways. In addition, I can record in Powerpoint slide by slide. Using Powerpoint on my Mac is also much less clunky that using Articulate Presenter in Windows on a Mac. It is especially easy when I need to make a small update, such as finding and fixing a typo in a slide, and not having to launch Windows, and then Articulate Presenter.

One can see an example of the results of my new approach in the image below from a Web site that I maintain, What is Biomedical and Health Informatics? I have maintained this site for many years, both as a way to introduce the world to my view of the field of informatics and also a showcase for the online learning modalities that our program uses. The site also has a great domain name I was able to snag when .health was made a top-level domain, http://informatics.health/. I have redone the lectures on this site using my new approach.

Certainly a bottom line is that the ease of updating lectures far outweigh the small additional features provided by Articulate Presenter. It may seem somewhat ironic that improvements in Microsoft Powerpoint now give me the ability to move on from Microsoft Windows and Articulate Presenter. But I am pleased that I can now can evolve with Apple as it innovates its hardware and software with its new ARM chips.

Annual Reflections at the End of 2021

As is the tradition with this blog, I end each year with a reflective look at the year past and what the future may hold. The year 2021 is not ending quite like I anticipated. At the beginning of the year, there were stirrings of optimism. The new COVID-19 vaccines had been released, and a new approach to political leadership had just been elected in Washington, DC. On January 2nd, I received my first dose of the Pfizer vaccine, followed by a second one three weeks later.

By the spring, the pandemic was seeming to wane. In late June, Oregon lifted all pandemic restrictions, and life seemed to be returning to relative normal. At work, my department began making plans for returning to the office, aiming to slowly transition over the summer and then return to a new state of normal when the regular academic year started in the fall.

 

Alas, it was not to be. By August, the Delta wave was surging and plans for returning to the office were postponed. Things became more optimistic into the fall, with my getting my booster vaccine in early October, but then, just when it seemed that the Delta wave was on the decline, the new Omicron variant emerged.

Nonetheless, there is some reason for optimism in the long run for the pandemic. The vaccines are having an impact, even if mostly among the vaccinated. Clearly the vaccines are not completely protective against infections, but they do appear to blunt the worst complications of the disease. In addition, there are new COVID-19 treatments becoming available in pill form, such as Paxlovid, which may provide a means to further prevent the worst impacts of infection. With my three doses of Pfizer vaccine on board, I feel relatively well protected, and feel confident that if I were to have a breakthrough infection, the course would be mild. (Though I am not taking any excessive chances.)

Although I look forward to getting back to working in person, I have managed to maintain my own productivity by working mostly from my home office. One thing that the pandemic has taught the informatics field, and certainly the work of my own department, is that we can pretty much function virtually. I still enjoy interacting in person with colleagues, but clearly most of the research and educational work we do can be carried out in a virtual manner.

 

I am sad that this pandemic has been so politicized. Every happening, often having nothing to do with the action of politicians, seems to require being viewed through a political lens. While there is, as with all science, some room for debate on translating scientific findings into public policy, it is unfortunate that many, especially those on the right, have chosen to frame everything in that context and make decisions that defy sound public health practice. I don't enjoy wearing masks or avoiding large gatherings more than anyone, but I recognize these as a small price to pay to protect public health. And it is sad to me that new vaccine technology, one of the greatest public health advances in human history, has also become a political Rorschach test.

The ability to function virtually would not be possible without the maturing of computing and network technology. It is still far from perfect, but we almost take for granted now that we can have a synchronous video call with almost any location on the planet. Social media has played a valuable role too, especially in my case, Facebook. While I am no big fan of Facebook's business model, and I would happily pay a modest subscription fee to get the benefits of it that I like, it has been invaluable in keeping up personally and professionally with friends and colleagues around the world.

In terms of the informatics field, I am optimistic that we will continue the contributions we make to health and larger society. While the pandemic has caused a detour in our priorities, we are still moving forward in core advances, particularly turning the potential of machine learning and artificial intelligence into reality and advancing standards to make data more interoperable. The pandemic has exposed the limits of our current health information systems and processes, and even if too late to provide optimal value for this pandemic, will hopefully put us in a position to benefit the next public health crises.

This blog itself continues to be successful. This is the 361st post, and it is closing in on 768,000 page views over its nearly 13 years. This blog has provided me an opportunity to provide commentary on many topics, mainly professionally but also personally, and is easily accessible from the Internet. My frequency of posting has decreased some over the years, but I still appreciate the opportunity to weigh in on important topics and not feel the need to post at any specific level of frequency.

I head into 2022 with the optimism that things will get better, yet also with the realism that unforeseen events will happen and set us back. The key, as always, is to live life in a way that one has no major regrets for paths not taken, and living in a way that is just and equitable for the rest of society.

From Reading to Writing: Next Steps for Patient Data Exchange and Interoperability

The rationale and implementation for reading data from the electronic health record (EHR) and other clinical sources is relatively simple and straightforward. Especially now enshrined into law in the US by the 21st Century Cures Rule, and standardized by the FHIR application programming interface (API), accessing data for reading by clinicians, patients, and others is here to stay.

Writing data to the EHR or other clinical information systems is a little more complicated. As in all aspects of informatics, the technology part is relatively simple, as activating the API in the reverse direction is not difficult technologically. But writing data into the EHR and other systems raises a number of issues. Earlier this year, the Office of the National Coordinator for Health IT (ONC) convened a workshop to address this topic. The workshop discussed stakeholder knowledge, current usage, potential use cases, and lessons learned on “write-back” API functionality. A report from the workshop was released in November, 2021 and provides excellent insights into the usage and challenges for such technology. Five categories of stakeholders were represented: researcher, technologist, healthcare provider, patient, and financial technologist. (The provider perspective was provided by OHSU faculty Dr. Ben Orwoll.)

The report summarized a number of possible use cases and their data sources for API write-back:

• Data from devices, such as wearables and remote monitors
• Questionnaires from patients or care activities
• Results of risk scores and calculators
• Patient input of symptoms or reported outcomes
• Recommendations of clinician decision support
• Annotation or amending of patient notes
• Results or recommendations of machine learning/artificial intelligence algorithms
• Data from transitions of care across orgniazations
• Community sources of data, including social determinants of health

Also identified were a number of technology barriers to writing data back into the EHR:

• Limitations of FHIR standard
• Accuracy and completeness of the data
• Security of data from third-party apps
• Mapping and coding issues for data entering EHR and other systems
• Patient-matching accuracy
• Requirements for manual workflow and/or reconciliation
• Obligations of organization and clinicians for data written back

In addition, the report raised a number of policy, preference, and data use concerns:

• Data ownership and expectations for patients and clinicians
• Compliance with HIPAA and other current laws
  
• Relationship to designated record set and legal medical record
• Regulation needed to support open APIs and their adoption
• Requirements for future policy

The entry of data into the EHR beyond the usual documentation by clinicians and others changes the nature of the EHR, marking the true transition from electronic medical record to electronic health record.

Kudos for the Informatics Professor - 2021 Update

Although virtually all work remains remote through the end of 2021, the Informatics Professor has still been productive in his work and actively collaborating with others around the world. I am proud of my accomplishments in 2021, and this post lists some of the highlights.

Probably my biggest highlight for this year is the awarding of two new National Institutes of Health (NIH) grants. One is an exciting new project funded by the NIH Harnessing Data Science for Health Discovery and Innovation in Africa Initiative (DS-I), where the Oregon Health & Science University (OHSU) Biomedical Informatics Graduate Program will be teaming up with the University of Cape Town (UCT) to develop new Research Training opportunities. A nice overview in Nature describes the larger initiative.

In this project, OHSU will partner with UCT to develop a new graduate program in Computational Omics and Biomedical Informatics (COBIP) that will start at UCT and aim to expand to other institutions from other countries across southern Africa. I will be on the leadership team and provide my expertise in developing and leading graduate programs in biomedical informatics and data science. The program will draw on the clinical informatics courses that OHSU offers, and other OHSU informatics faculty also be involved in teaching and mentoring of research projects of students in the program. We will aim to build what the overall NIH program seeks to accomplish, namely adding capacity in biomedical informatics and data science across the African continent. With the development of computer networks and sources of health data across Africa, the time is ripe to leverage it for the health of people in Africa. As with other programs we have helped developed over the years (in places like Argentina and Thailand), our goal is to develop sustainable programs in these places so they can eventually function without the need for us. Despite some setbacks from Covid, Africa is really poised to see development of using informatics to improve the healthcare and public health systems, ultimately benefiting health of the people. And hopefully in the long run the program will attract local talent and expertise and become an independent part of the global informatics community.

The second new grant is a five-year renewal of a project, Semi-structured Information Retrieval in Clinical Text for Cohort Identification. This funding continues my work in collaborating with colleagues from the Mayo Clinic and University of Texas Houston at the intersection of information retrieval and re-use of electronic health record data.

In addition to grants funded, I also served as a senior author on  a couple of journal articles published in 2021:

• Chen J, Hersh W, A comparative analysis of system features used in the TREC-COVID information retrieval challenge, Journal of Biomedical Informatics, 2021, 117:103745.
• Roberts K, Alam T, Bedrick S, Demner-Fushman S, Lo S, Soboroff I, Voorhees E, Wang LL, Hersh WR, Searching for answers in a pandemic: an overview of TREC-COVID, Journal of Biomedical Informatics, 2021, 121:103865.

In 2021, I also assumed the Presidency of the International Academy of Health Sciences Informatics and was a co-author on two papers describing some activities of the Academy:

• Koch S, Hersh WR, Bellazzi R, Leong TY, Yedaly M, Al-Shorbaji N. Digital Health during COVID-19: Informatics Dialogue with the World Health Organization. Yearbook of Medical Informatics. 2021, 30(1):13-16.
• Haux R, Ball MJ, Hersh WR, Huesing E, Kimura M, Koch S, Martin-Sanchez F, Otero P. The International Academy of Health Sciences Informatics (IAHSI): 2020 Report. Yearbook of Medical Informatics. 2021, 30(1):8-12.

I also published three book chapters in 2021:

• Hersh W, Biomedical Informatics, in Kutz M (ed.), Biomedical Engineering Fundamentals, Third Edition, McGraw-Hill, 2021, 31-48.
• Hersh W, Information Retrieval, in Shortliffe EH, Cimino J, Chiang MF (eds.), Biomedical Informatics: Computer Applications in Health Care and Biomedicine, 5th Edition, New York, Springer, 2021, 761-800.
• Hersh W, A Passion and a Calling, in: Kulikowski, C., Mihalas, G., Yacubsohn, Y., Greenes, R., Park, H.-A. (eds.), IMIA History Book. Healthcare Computing & Communications Canada, 2021, 383–386 (described further in an earlier post to this blog).

During the year, I also participated, mostly remotely, in a number of academic conferences, some of which were international. I gave an international keynote talk, although unfortunately from the confines of my home office at not with my long-time friends in Argentina, at the Jornadas de Informática en Salud del Hospital Italiano de Buenos Aires on November 16th entitled, Artificial Intelligence in Medicine: The Need to Translate From Basic Science to Clinical Value. Both the video part of my talk and my slides and references can be viewed online.

In another international conference in Portugal, I spoke virtually on a panel at the conference, Healthcare in Post-Pandemic – Will Value for Health be the Way?, hosted by the Value for Health Collaborative in Lisbon, Portugal (virtual), June 18th. Domestically, I provided an update virtually, AMIA Update on Academic, Education and Certification Initiatives, at the CAHIIM Fall Festival: Navigating Higher Education 2021 (virtual), October 7th.

Fortunately, I was able to participate in one conference in person. After a year and a half of not attending any conference in person, I was able to make it to the AMIA Annual Symposium in San Diego, CA from October 29-November 3. This meeting marked my 36th consecutive year attending the AMIA Annual Symposium, which was known as the Symposium on Computer Applications in Medical Care (SCAMC) when I first attended it in 1986. Last year, of course, the meeting was completely virtual, but fortunately the meeting took place in San Diego, where the mild climate of southern California allowed all social events to be outdoors. At the conference I participated in a panel entitled, Career Development Issues for Women in Biomedical Informatics Within Professional Organizations, on November 1st.

I was also interviewed for a couple of podcasts. One was, Conversations from the AMIA 2021 Symposium, from The Public Health Networker podcast of the Public Health Podcast Network, in which I gave some history and perspective of AMIA and its Annual Symposium over the years.

Another podcast in which I took part was from the series called Sound Practice, hosted by the American Association for Physician Leadership. My session was At the Intersection of Technology and Healing: Trends in Medical Informatics.

Finally, I also had a chance to do some writing focused on local issues in Portland, OR and the cities challenges in grappling with economic and social issues exacerbated the pandemic. I authored a perspective for a series in the Portland Business Journal entitled, Don't Count Portland Out (apologies for it being behind a paywall).